Most people understand that multitasking human effort is inefficient. This week, Payson explains why this is so in business terms that can be used to appeal to managers and executives, who sometimes request multitasking without knowing the consequences.
Multitasking-performing multiple tasks at the same time or, more accurately, time-slicing between tasks so that they execute concurrently-was a profound innovation for computer systems but doesn't work well for human activities. Human brains do not multitask well. When we shift our attention from one activity to another, there is a loss of productivity as we try to recall where we were when we return to a task in progress. You have experienced this when you were reading something dense and were interrupted by a phone call. When you return to your reading, it takes a moment to recall and re-create the context. This "restart time" is unavoidable and represents lost productivity and wasted time.
Most of us intuitively recognize the inefficiency of multitasking, particularly for complex, thought-intensive tasks. Imagine trying to write an opera, design a complex data system, and do your taxes while shifting your attention from one task to another every ten minutes.
If multitasking is inefficient, why do so many project plans call for people to work concurrently on multiple tasks? There are a few good reasons:
- Allocation of senior staff, who provide specialized consulting to a task on a part-time basis to coach or assist junior staff, who are doing the heavy lifting.
- Tasks that are inherently "bursty," requiring focused effort interspersed with delays. Imagine a task to staff a help desk-it needs you when the phone rings, but you may have available time when things are quiet. Another low-priority task could run concurrently to maximize resource usage.
- I could only think of two. Anyone else have good reasons? If so, comment below and let me know.
The list of good reasons is short. When I review projects that have people "multitasked," it often isn't about the nature of the work. It is a thin veil trying to cover the project's not having sufficient staff. Rather than admitting that there aren't enough people to get the work done, multiple tasks are assigned to the same people and scheduled to run concurrently. Later, everyone acts surprised at missed delivery dates, and managers rattle their sabers and makes noises about "holding people accountable" for their schedules.
Imagine I am asked to estimate tasks A, B, and C, each of which produces a needed work product. Suppose I respond that each task will take a week of my time. Do you imagine that through the magic of concurrent scheduling I could get all tasks done in the same week? If I had assumed that each task was a full-time effort, then the initial estimate was for three weeks of my effort. If we run the tasks concurrently, does it become less work? No, actually it becomes MORE work because of the inefficiency of task switching. Rather than three weeks' worth of effort, it will require three weeks plus whatever the overhead factor is for task switching. Yet, despite evidence of its inefficiency, some managers can't resist the pressure to schedule tasks concurrently, even if it means elongating the schedule for all three tasks.
A few years ago, a friend named Dohn showed me a simple picture that explains why even this kind of multitasking is often a bad business decision. Let's look at three tasks run concurrently.
Even if we imagine that we can context switch with perfect efficiency, when is the first work product received? At the end of Week 3, with the other two tasks' work products.
If I get the sniffles at the end of the second week and must go to a hospital, what do you have to show for my labor? Nothing.
Contrast that with work scheduled without concurrency.
When is the first work product received? If I get the sniffles at the end of the second week, what do you have to show for my labor?
Which schedule is the lowest risk?
Multitasking is often a poor business decision. It is inherently inefficient and defers realizing the benefit from work performed while increasing risk, often with little or no benefit. If this argument can help reduce the amount of multitasking your managers request, that would be great. If it can't help you with your managers in the short term, I hope that, when you are in a position to influence how tasks are assigned, you might be encouraged to resist the temptation to multitask except where someone can make a compelling business case in favor of it.